104 VARIABILITY 



in some measure at least, by the environment, for such tumours 

 are seen, especially in irritated tissues. Germ-cells, or cells of the 

 germ-tract, appear also to be affected sometimes by the direct 

 action of the environment, as in the case of Clayton's beans and 

 European dogs in India. But, as we have insisted, evolution is 

 never perfect ; the ' laws ' created by it have ever exceptions ; 

 for it is only by eliminating the exceptions that nature creates 

 and maintains uniformities amongst living beings. The fact that 

 these exceptions occur is in accordance, therefore, with all we 

 know concerning the adaptations of living beings. 



169. The advantage of the practical limitation of the occur- 

 rence of variations to a particular part of the germ-tract is 

 obvious. If variations occurred amongst somatic cells, they would 

 tend to ruin co-adaptation among the cells that varied (and their 

 descendants) and the other somatic cells (and their descendants) a 

 co-adaptation which has been achieved only by prolonged evolution. 

 A cancer which results, seemingly, from a variation in a somatic cell 

 (a variation, like other variations, inheritable by the descendants of 

 the cell) is an example. Moreover, such variations, even if favourable 

 to the individual, would be useless to the species, for, save in the 

 event of the Lamarckian hypothesis being true, they would not be 

 reproduced by the offspring of the individual. On the other hand, 

 germinal variations, while affording the necessary materials for 

 Natural Selection, do not affect the well-being of the community of 

 which the germs are members, for here the germs have no functions. 

 Again, if variations occurred all along the germ-tract, they would 

 accumulate, and offspring would tend to vary, not in the right 

 degree, but in proportion to the length of the tract and in pro- 

 portion to the duration of time that clasped between the fertilizing 

 of the ovum from which the parent arises and the production of 

 his offspring ; that is, the offspring earliest produced would tend 

 to vary less than those produced later. Lastly, if the point at 

 which variations normally occur were situated, not at the end of 

 the germ-tract, but earlier in it, then, in proportion as the point 

 was near the beginning and distant from the end of the tract, 

 larger and larger batches of germ-cells would, as I say, tend to be 

 identical. By limiting normal variability to the end of the germ- 

 tract, nature ensures, on the one hand, that all germ-cells shall vary 

 from one another, and, on the other, secures control over variations, 

 which are then soonest put to the test of fitness to the environment. 



170. It may be thought that I have carried deductive thinking 

 too far in the preceding paragraphs. But the facts on which I 



